Tunable diode laser spectroscopy and line parameters of water vapor in the near-infrared

JOURNAL OF SCIENCE OF HNUE Mathematical and Physical Sci., 2014, Vol. 59, No. 7, pp. 157-164 This paper is available online at TUNABLE DIODE LASER SPECTROSCOPY AND LINE PARAMETERS OF WATER VAPOR IN THE NEAR-INFRARED Ngo Ngoc Hoa and Nguyen Manh Nghia Faculty of Physics, Hanoi National University of Education Abstract. In this study, we present the technique of tunable diode laser spectroscopy and the spectroscopic parameters of 13 lines of H2O in the 11980 -12260 cm−1 spectral region. Spectra were recorded at room temperature for a wide range of pressures (2 - 15 Torr for pure H2O and 50 - 760 Torr for H2O in air). Line parameters were adjusted from experiments using the least square method, Fortran and three line-shape models: the Voigt profile (VP), the (hard collision) Rautian profile (RP) and the speed dependent Voigt profile (SDVP). A comparison between our results with the HITRAN database 2012 shows good agreement. Keywords:Water vapor, line-shape, tunable diode laser spectroscopy. 1. Introduction Water vapor is a key molecule in the Earth’s atmosphere, its distribution the object of several remote sensing experiments. The latter provide recordings of atmospheric absorption spectra whose inversion yields the atmospheric humidity vertical profile. For such applications, precise knowledge of the spectroscopic parameters of the H2O lines is needed. In this work, line parameters (position, integrated intensity and broadening coefficient) are deduced from laboratory spectra assuming a specific line-shape. For this, the Voigt profile [1] is used in most of the available studies. Within this model, two collisional parameters are used, i.e. the Lorentz broadening and shifting coefficients, the Doppler width being fixed to its theoretical value. As is well known, [2-4], the Voigt model can lead to discrepancies with measured spectra since it does not take into account two velocity effects: (a) collision-induced velocity changes, leading to the so-called Dicke narrowing and (b) the speed dependence of the pressure-induced width and shift. Using more refined models that take into account such effects, several studies showed that the line broadening determined by the Voigt profile can be underestimated up to 10% [2, 4] while the error for integrated intensity is from 0.3 to 2% [5, 6]. Received August 12, 2014. Accepted October 13, 2014. Contact Ngo Ngoc Hoa, e-mail address: hoa.nn@hnue.edu.vn 157 Ngo Ngoc Hoa and Nguyen Manh Nghia Three spectral shape models have been used: the Voigt, the Rautian (to take into account the Dicke narrowing effect) and the speed dependent Voigt models (to take into account the speed dependence of the collisional parameters). The corresponding absorption coefficients for a single line are given by [7]: αV (σ) = S.PH2O√ π √ ln2 ΓD Re {W (σ − σ0,ΓD,∆,Γ)} , (1.1) αRP (σ) = S.PH2O√ π √ ln2 ΓD Re { W (σ − σ0,ΓD,∆,Γ + B) 1−√πBW (σ − σ0,ΓD,∆,Γ + B) } , (1.2) αSDV P (σ) = S.PH2O π3/2 √ ln2 ΓD Im  +∞ ∫ −∞ e−t 2 [1− Z ′(t)] (σ − σ0) √ ln2 ΓD − t− Z(t) dt  , (1.3) where PH2O is the partial pressure of H2O and S is the integrated intensity of the line. σ0 and ΓD are the unperturbed spectral position of the transition, and the Doppler width Γ, △ and B are the collisional half-width (HWHM), the pressure induced line-shift, and the narrowing parameter, respectively. The complex probability function W is given by: W (σ − σ0,ΓD,∆,Γ) = i π +∞ ∫ −∞ e−t 2 (σ − σ0 −△) √ ln2 ΓD − t+ iΓ √ ln2 ΓD dt. (1.4) The function Z(t) and its derivative Z’(t) in equation (3) are given by: Z (t) = √ ln2 ΓD [∆ (v˜t) + Γ (v˜t)] , (1.5) Z ′ (t) = v˜ √ ln2 ΓD [ ∆ ′ (v˜t) + Γ ′ (v˜t) ] , (1.6) where v˜ is the most probable speed, △ (v˜t) and Γ(v˜t) being the speed dependent line shifting and broadening parameters. The speed dependence of the line broadening is modeled using a quadratic function [8]. 2. Content 2.1. Experimental setup The experimental setup used for the measurements of spectroscopic parameters of water vapor in the near-infrared spectral region is shown in Figure 1. An external-cavity 158 Tunable diode laser spectroscopy and line parameters of water vapor in the near-infrared diode laser [9] provides a wide tuning range of about 850 nm at room temperature. The diode laser wavelength can be continuously scanned by means of a voltage-ramp signal that simultaneously drives the laser injection current and the piezoelectric transducer that sets the angle of the external-cavity grating. The emission line width of the diode laser is about 1 MHz. In this spectral region, we investigated lines of the 2ν1 + ν2 + ν3 bands of H2 16O by scanning intervals about 0.8 cm−1. Figure 1. Experimental setup used for the measurements of line parameters in the near-infrared region A White blood cell with a total optical path of 40 m was used. The experiments were carried out as follows: a spectrum was first recorded with the empty cell, providing the 100% transmission reference. The cell was then filled with H2O vapor at pressures ranging from 2 to 15 Torr, well below the saturated vapor pressure (about 18 Torr at room temperature) and self-broadening spectra were recorded. For air-broadening spectra, the cell was first filled with pure H2O (from 4 to 7 Torr) before ambient air was added up to total pressures of 50 to 760 Torr. Partial pressures of H2O in the mixtures were 159 Ngo Ngoc Hoa and Nguyen Manh Nghia then calculated using the integrated intensity determined from the pure H2O spectra. All measurements in this study are realized by diode laser system at Paris East University. A Fabry-Perot cavity was used for the calibration of the wavenumber scale. Two spherical mirrors have a curvature radius of 75 mm, resulting in a free spectral range of 1 GHz (Figure 2). The finesse is 1000 leading to a resolution of 1 MHz. The peak positions of this Fabry-Perot etalon allow a determination of the relative frequency scale with a precision of better than 1 MHz (or 3.3.10−5 cm−1) while scanning the diode laser. The peak positions of the Fabry Perot cavity were calibrated to relative wave numbers by using a third degree polynomial function. Figure 2. An example of the signal by recording simultaneously the detector signal, the current ramp and the Fabry-Perot signal 2.2. Results and discussions The measured spectra were least-squares fitted with the three profiles described above and Doppler widths fixed to theoretical values. The integrated intensity, the effective line position, and the line-shape parameters (Γ, B, Γ2), together with two parameters describing the zero absorption level, are adjusted in the fits. As expected and exemplified by Figure 3, the VP leads to significant residuals, and the RP and SDVP are in better agreement with measured spectra. The three lowest panels show the differences between measured absorptions and those adjusted using the VP, RP and SDVP. 160 Tunable diode laser spectroscopy and line parameters of water vapor in the near-infrared Figure 3. Room temperature measured absorptions of the 606 ← 505 (12249.3895 cm−1) transition of pure water vapor (left) and of H2O in air (right) at different pressures 2.2.1. Integrate intensities For the integrated intensity, only the results from pure H2O spectra were used. As can be seen from this table and Figure 4, the RP and the SDVP lead to very close values, both slightly larger (typically by 0.8%) than those determined with the VP. This result is fully consistent with those obtained in other spectral regions [5, 6]. Figure 4. Ratios of the intensities of the 13 considered transitions obtained by the RP (◦) and SDVP () to those obtained by the VP 2.2.2. Broadening coefficients The self-broadening coefficients were determined from fits of pure H2O spectra, shown in Figure 5a, display the collisional line-widths of the transition at 12249.3895 cm−1, which shows that they have the expected linear variation with the H2O pressure. The slopes give broadening coefficients. For example with the transition at 12249.3895 cm−1, the corresponding broadening coefficients obtained with the VP, RP and SDVP are 161 Ngo Ngoc Hoa and Nguyen Manh Nghia 0.429, 0.446 and 0.452 cm−1.atm−1, respectively. The smaller value obtained with the VP shows the effect of the line narrowing, observed in Figure 3. For spectra of H2O diluted in ambient air, the total line-width of a transition is written as: Γ = γself .PH2O + γair. (P − PH2O) , where γself and γair are the self and air-broadening coefficients, respectively. P is the total pressure in the celle. In order to determine γair, the partial pressure PH2O in the cell was for each spectrum first determined from the ratio of the integrated area under the measured spectrum (obtained from fits) to the line intensity determined from the spectrum of pure H2O. A linear fit of (Γ− γself.PH2O) versus Pair = (P - PH2O) gives γair. As exemplified in Figure 5b, the values obtained vary linearly with Pair. The slopes from Figure 5b (0.0763, 0.0792 and 0.0799 cm−1.atm−1 obtained with the VP, RP and SDVP, respectively) again demonstrate the consistency of the RP and SDVP determinations and the underestimation of the broadening by the VP of about 5% as presented in Figure 6. Figure 5. Line widths of the transition at 12249.3895 cm−1 as functions of (a) the H2O pressure and (b) air pressure, obtained using the VP, RP and SDVP Figure 6. Ratios of (a) the self and (b) the air-broadening coefficients of the 13 considered transitions obtained using the SDVP () and the RP(◦) to those obtained using the VP 162 Tunable diode laser spectroscopy and line parameters of water vapor in the near-infrared The H2O vapor self- and air-broadening coefficients from this study obtained using the VP are compared with data from the literature HITRAN08 [10] and HITRAN12 [11]. Note that our results are in good agreement with all previous studies as shown in Table 1 and Figure 7. The self-broadening coefficients from this study are slightly lower than the values in the databases with a mean deviation of 2% and 7%. However, the air-broadening coefficients are about 4% higher than the values in the databases. Table 1. H2O self and air-broadening coefficients at 296 K S, 0, cm−1 10−23 cm−1 molec−11 cm2 self, cm −1 atm−1 air, cm−1 atm−1 VP VP [10] [11] VP [10] [11] 11988.4939 1.0626 0.369 0.353 0.352 0.0681 0.0708 0.0708 11988.7257 0.3422 0.314 0.364 0.352 0.0654 0.0742 0.0742 12226.1012 4.8473 0.526 0.461 0.483 0.0931 0.0933 0.0954 12236.5601 4.0128 0.475 0.416 0.412 0.0850 0.0842 0.0849 12244.7186 3.5084 0.467 0.424 0.398 0.0899 0.0952 0.0958 12244.7881 1.0197 0.392 0.299 0.402 0.0694 0.0681 0.0681 12248.5787 0.9492 0.362 0.352 0.391 0.0703 0.0754 0.0774 12249.3895 2.6468 0.429 0.376 0.391 0.0763 0.0773 0.0776 12259.4776 0.5645 0.354 0.323 0.313 0.0665 0.0691 0.0673 12259.6033 1.6060 0.358 0.330 0.348 0.0627 0.0621 0.0621 12268.9969 1.0111 0.300 0.258 0.352 0.0530 0.0618 0.0618 12280.4387 0.2820 0.389 0.429 0.421 0.0754 0.0801 0.0801 12280.6634 0.9262 0.374 0.329 0.343 0.0719 0.0725 0.0725 Figure 7. A comparison of the H2O self- and air-broadening parameter from the present work with those from other studies [10, 11] as a function of line intensity 163 Ngo Ngoc Hoa and Nguyen Manh Nghia 3. Conclusion Absorption spectra of 13 lines of H2O in the near infrared spectral were recorded at room temperature using a tunable diode laser system. Line parameters were determined from experiments using three line-shape models: the Voigt profile (VP), the (hard collision) Rautian profile (RP) and the speed dependent Voigt profile (SDVP). The results show that the RP and SDVP are in better agreement with measurements than the VP and that they lead to larger values of the line parameters (about 5% for the line broadening and 0.8% for line intensity). The comparison between our results and the HITRAN database 2008 and 2012 shows good agreement. REFERENCES [1] W. Voigt, 1912. U¨ber das gesetz intensita¨tsverteilung innerhalb der linien eines gasspektrams. Sitzber. Bayr Akad., Mu¨nchen Ber., pp. 603-620. [2] D. Lisak, J. T. Hodges and R. Ciuryło, 2006. Comparison of semiclassical line-shape models to rovibrational H2O spectra measured by frequency-stabilized cavity ring-down spectroscopy. Physical Review, A 73 012507. [3] H. Tran, D. Bermejo, J. L. Domenech, P. Joubert, R. R. Gamache and J. M. Hartmann, 2007.Collisional parameters of H2O lines: Velocity effects on the line-shape. Journal of Quantitative Spectroscopy and Radiative Transfer, 108 pp. 126-145. [4] M. D. De Vizia, F. Rohart, A. Castrillo, E. Fasci, L. Moretti, L. Gianfrani, 2011. Investigation on speed dependent effects in the near-IR spectrum of self-colliding H2 18O molecules. Physical Review, A 83, 052506. [5] H. Li, A. Farooq, J. B. Jeffries, R. K. Hanson. Diode laser measurements of temperature-dependent collisional-narrowing and broadening parameters of Ar-perturbed H2O transitions at 1391.7 and 1397.8 nm. Journal of Quantitative Spectroscopy and Radiative Transfer, 109, pp. 132-143. [6] D. Lisak, J. T. Hodges, 2008. Low-uncertainty H2O line intensities for the 930 nm region. Journal of Molecular Spectroscopy, 249, pp. 6-13. [7] J. M. Hartmann, C. Boulet and D. Robert, 2008. Collisional effects on molecular spectra. Laboratory experiments and models, consequences for applications, Elsevier. [8] F. Rohart, H. Mader and H. W. Nicolaisen, 1994. Speed dependence of rotational relaxation induced by foreign gas collisions: Studies on CH3F by millimeter wave coherent transients. The Journal of Chemical Physics, 101, pp. 6475-6486. [9] Toptica Photonics. Diode Laser Series: DL 100, User Manual, [10] L. S. Rothman, I. E. Gordon, A. Barbe, D.C. Benner, P. F. Bernath et al., 2009. The HITRAN 2008 molecular spectroscopic database. Journal of Quantitative Spectroscopy and Radiative Transfer, 110, pp. 533-572. [11] L. S. Rothman, I. E. Gordon, A. Barbe, D.C. Benner, P. F. Bernath et al., 2013. The HITRAN 2012 molecular spectroscopic database. Journal of Quantitative Spectroscopy and Radiative Transfer, 130, pp. 4-50. 164